A recent report on the songs of the eponymous “great tit”, a common forest bird famous for learning to peck the foil tops of milk bottles in the 1950s, shows that they independently acquire a deeper song when in urban environments than when in forest environments. As the writer at ScienceNOW tells it, in forests they sound like Barry White, and in cities like Michael Jackson.
Passerine songs are usually adapted to the acoustics of their usual environment. Birds in denser vegetation will, I am told, end their songs on a rising sharp note, because there is more absorption of sound than in an open forest or field, and so it has to carry to assert territory or find mates and flockmates.
What’s interesting here is that these birds are culturally adapting to the same environment in different places. Hence they are fitter, due to individual learning, but, and this is the point I want to make, so too are the songs themselves. A deep Barry White lerve song will get passed on more effectively if it works better in a city, than it will in the forest. In addition to tracking the fitnesses of the organisms that learn the song, we can also track the fitnesses of the song. And since songs can be spread independently of genes, these two fitnesses are not the same.
I want to continue my presentation of science as evolution, and draw a parallel here with the fitnesses of scientific ideas, norms, and practices, which we will collectively call “memes” on the understanding that nobody at any time is to challenge me about what the hell a “meme” is.
Scientific memes are generated – how is the province of cognitive psychology and indeed pathological psychiatry sometimes – and transmitted by word, deed (showing grad students and visiting researchers), and publication. Some of these are taken up and cited. Some are cited very highly. Others may be cited once, or, in the case of the majority of papers, never cited and rarely read at all. What, we must ask, determines what gets passed on and copied by other researchers, and what doesn’t?
The most obvious answer is: empirical data. However, that cannot be a full answer, not least because empirical data is, as philosophers of science have been saying since the 1950s, highly theoretical in nature – what data gets measured, what techniques and equipment gets used, and what analytic techniques go to make up a reported data set is dependent largely on what is the prevailing set of conventions, protocols, norms and generally accepted behaviours in the discipline. Hence, data is itself part of the evolutionary process.
Of course, this is equally true of selective pressures in biology – apart from abiotic resources like solar energy and geography, organisms are often adapting mostly to each other – predator to prey and the reverse, as well as communities that act as wholes when it comes to constructing their environments, and of course organisms adapting to their conspecifics. This is not, therefore, a dysanalogy between biological and cultural evolution; it pays not to expect the simple answers to be right in this matter.
Popper, of course, flew a kite about “verisimilitude”, in which theories approached truthlikeness as they survived increasingly strong selection pressures by falsification tests. Nothing, so far as I can tell, involves truth in any way when talking about a simple evolutionary process, since truth and fitness are distinct topics. A false idea may survive indefinitely in science if it has a high fitness, just as our psychology can impose false ideas about the world as a result of selection.
One of the fitness-defining influences, then, on scientific memes has to be the other researchers in the field. If they consider a meme worth reproducing, for whatever reason, they will reproduce it. If they cite the originator of the meme, then that researcher’s “conceptual inclusive fitness”, as Hull calls it, is enhanced. If not, the fitness of the meme can still be high in its own right.
Some have taken the social-relativity of scientific memes (by other names) to mean that science is entirely socially constructed. But this is not the case – a lot of data is not theory-dependent, or, if it is, the theory is very low level (this microscope works, this technique is an engineering tool) rather than a disconnected high level Weltanschauung or “paradigm”. A lot of the phenomena on which science is based are very independent of theory. Facts may be hard things to cash out conceptually, but as far as I’m concerned, much of them are just observations by people familiar enough with the domain under investigation that they can just see them (species are phenomenal objects, in my opinion, as an instance of this). So science as a whole is grounded in non-theoretical phenomena. But once you move past the ordinary phenomenal empirical level, it gets harder to maintain that data is the sole selective pressure on scientific memes.
One of the innovations in Hull’s account of science is that science is organised, or rather self-organises, into cooperative “demes”, which raises the possibility that demic research programs are selected at a higher level than the conceptual fitness of individual scientists. Within the deme, selection is relaxed while a program sorts itself out and is elaborated. Between demes, such as competing labs, or funded programs, selection is intense, and so novel ideas are able to be buttressed against harsh attack until they can stand on their own legs in the wider community.
This means that some selection, for compatibility with other shared values within the deme, will go on before the idea is tested, although of course researchers will attempt to protect the idea by suggesting likely counterfactuals from other programs. This selection will occur in small groups, or even in a single individual head. The point of the demic structure of science is that it is in virtue of this very structure that science makes progress. Ideas that survive the friendly fire are more likely to get used by those outside the deme if they survive competitive attack.
One thing about this meme-centric view of scientific evolution, apart from the confusion of entities doing the evolving (is it a population of scientists, ideas, practices, labs, programs) that bothers me, is that it tends to be panadaptationist. Not everything that survives in science is a good idea. This, too, in necessary, for if all ideas were strictly constrained, say by eliminating anything that counts as non-science in the first instance, selection would tend to eliminate all sources of inspiration more rapidly than novel ideas arise. Think of true novelty as the analogue to mutation. Most new ideas are simply recombination of hidden or low frequency existing ideas. Einstein’s relativity was itself a matter of taking existing notions and turning them on their heads. Inspiration more often comes from trying out variations of existing themes than generating an entirely novel one. Nearly all the elements of Darwin’s theory were in play before him – the novelty came in combining them into a package with implications that species evolve, rather than staying static, and that relationships were a tree, not a network. If we eliminate all ideas that might be science by selection, such combinations may be impossible later on. There’s an analogy here with sex and advantageous mutations.
So it is important that there be some slack in how scientific ideas are dealt with at various levels, and it also is important that ideas in the wider culture can inspire and sometimes constrain scientific ideas. There is a boundary, of course, between culture and science, just as there is between other aspects of human activity and the wider culture (such as theology, art, sport, and political theory, to take a bunch of domains at random), even though they all are influenced by the wider society. In our evolutionary-genetic analogy, memes have to be able to drift, not too much or the scientific enterprise will evaporate, and not too little, or the disciplines will end up stuck on local fitness peaks, which may be determined more by internecine politics than by actual investigation.
Seeing science as a selection process (plus drift) moves us away from such questions as “what is the scientific method?” There is no single method that science has that is essential to it. Methods come and go; and their usefulness is the fitness they add to the scientific activities of the players. A professional career for a scientist is determined by the strategies that researcher adopts – one may be an early adopter of a novel and untested idea – high risk, high payoff if correct – or a late adopter of conventionally accepted ideas – low risk, low payoff – or any of a slew of intermediate and mixed strategies. Overall, though, these strategies are themselves the outcome of selection for successful return on investment, as it were. We should expect that scientists overall will tend to cluster about the “best” strategy for advancing their careers, and hence, inadvertently advancing their discipline.
Just like our passerines, the song of the scientist will adapt to be heard over the hubbub of both their peers and their social environment. When social intervention is low, the science will tend to be driven by internal interactions and empirical engagement (but not always – there are theoretical disciplines too); when it is high, it will tend to evolve along with the societal norms. The latter can be devastating for a science. When governments have regulated and directed, through funding control, the sort of science that is done, the end result is almost always a moribund outcome (consider the “war on cancer” in the Nixon era, which was devoted to the idea that all cancer was viral). When societal intervention is too low, scientists can tend to wander off in sometimes nasty ways. Like everything else, there is a happy medium, but no general rules for what that might be. We only know what the fitness peak is after the event. Rules of thumb notwithstanding, science evolving is like every other evolutionary process – you can describe it, but not prescribe it.
Now I have an image of Barry White in a lab. I must go and cleanse my brain…